Unsupervised Domain Adaptation through Self-Supervision

Abstract

This paper addresses unsupervised domain adaptation, the setting where labeled training data is available on a source domain, but the goal is to have good performance on a target domain with only unlabeled data. Like much of previous work, we seek to align the learned representations of the source and target domains while preserving discriminability. The way we accomplish alignment is by learning to perform auxiliary self-supervised task(s) on both domains simultaneously. Each self-supervised task brings the two domains closer together along the direction relevant to that task. Training this jointly with the main task classifier on the source domain is shown to successfully generalize to the unlabeled target domain. The presented objective is straightforward to implement and easy to optimize. We achieve state-of-the-art results on four out of seven standard benchmarks, and competitive results on segmentation adaptation. We also demonstrate that our method composes well with another popular pixel-level adaptation method.

References in corpus (7)

• Rethinking Atrous Convolution for Semantic Image Segmentation
• Understanding deep learning requires rethinking generalization
• CyCADA: Cycle-Consistent Adversarial Domain Adaptation
• Domain Separation Networks
• Unsupervised Cross-Domain Image Generation
• Domain Adaptation for Visual Applications: A Comprehensive Survey
• Boosting Supervision with Self-Supervision for Few-shot Learning

Cited by in corpus (6)

• Self-supervised Knowledge Distillation for Few-shot Learning
• Rethinking Distributional Matching Based Domain Adaptation
• A Review of Single-Source Deep Unsupervised Visual Domain Adaptation
• Auxiliary Signal-Guided Knowledge Encoder-Decoder for Medical Report Generation
• Learning from Extrinsic and Intrinsic Supervisions for Domain Generalization
• Curriculum Manager for Source Selection in Multi-Source Domain Adaptation